I think I’m currently in the lead; so let the games begin. What numbers are the most expensive? The UCC firm prefix is around $150/year and about $750 get into the game. Martin Geddes takes a stab at the value of a phone number treating it as a license to bill customers he get’s a valuation of around $140 dollar a year.

I’ve become increasingly interested in the number business; i.e. selling minted registered official numbers. The raw material is free and plentiful. It is almost the canonical example of the commons. But as the examples above make clear, registered numbers are can be quite expensive. The act of converting a common number into private property is the business of a registry. The domain over which the numbers are used form a club, the registry plays a role in the regulation or governance of that club. That registry can be a private rent seeking entity, like Verisign, or a nonprofit that seeks less naively quantifiable goals like the UCC or ICANN.

Frame it how ever you like, there is always an owner (or owners) of the registry function. A bit of paranoia about these registries is in order. On the one hand they are well positioned to become abusive monopolies (or oligarchies) , while on the other they can suffer all the classic breakdowns that trouble commons.

The rent seeking registry owner strives to make the numbers scarce. To convince customers that common numbers are worth something are better than common numbers is a bit of a magic trick. The trick is one of illusion, faith, and fact.

Consider the prime numbers needed to implement https connections; i.e. SSL keys. How much of the value of a SSL key from a leading vendor is illusion, faith, or fact?

What concerns me is the lack of awareness among system designers of the choices they are making when they block out the design of these registries. Some design choices favor the emergence of a highly concentrated registrar market while others favor the emergence of a dysfunctional diffuse set of registrars. While either outcome has unfortunate side effects it appears to me that most designers aren’t even aware they are making these choices.

Consider an example – domain name registries. Rent seeking is common. Long tedious efforts by members of the club to reduce the degree of market concentration continue. I have been amused to notice a nice example of how making a market more diffuse creates challenges in market’s social contact. Consider this example: the scarcity of domain names provides the foundation for a lot of spam filtering and as domain names become less scarce those techniques are breaking down. Notice that these registry numbers provide the hook on which reputation hangs. A registrar that retains ownership of the number, and only licenses it’s use, can engage in bait and switch pricing, so I wonder about the recent offer of free domain names from a number of top level registrars. I don’t recall anybody who was aware of that this kind of game was being set up when the original designs for DNS were blocked out. Certainly some people were aware of the risk of having a single point of failure; but that’s just one example of something to be paranoid about.

Actually I don’t think it’s a necessary to have the contest. I think I know what the most expensive registered numbers are: citizen ID numbers. Those in wealthy nations with strong well functioning social contracts and deep pools of public services. A Manhattan phone number is a just proxy for a set of those. Which brings us full circle back to Martian’s posting about phone number values. Should the citizens of Manhattan demand that their phone numbers not be handed out to others; clearly that erodes the value of their numbers?

[[ My current favorite example of a light weight registrar: Linux User Numbers. ]]